Monitoring and Management for Metabolic Side Effects of Antipsychotics: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Monitoring and Management for Metabolic Side Effects of Antipsychotics (NCT01875861), a randomized na study. The primary outcome is Change in Site-level Rates of Weight Monitoring at Baseline (Within 30 Days of a New Antipsychotic Prescription).

2. Clinical Question

3. Trial Design

The registry describes a randomized, parallel-group study with no masking. The intervention was Evidence-Based Quality Improvement plus external facilitation; the comparison was usual care in the context of the MIAMI Project. The listed enrollment is 12. The registry record does not say whether that enrollment count refers to sites, individuals, or another unit, so it should not be interpreted as the number of patients contributing to the weight-monitoring outcome.

The study was intended to test an approach for implementing recommendations to assess and manage metabolic side effects of antipsychotic medications, including weight gain, diabetes, and elevated lipids. The primary measure focuses on weight monitoring, rather than measuring all of these metabolic effects. The record describes repeated monthly measurement through pre-implementation, implementation, and sustainability phases. It does not specify the precise duration of each phase beyond describing a 6-month pre-implementation phase and referring to 6-month phases.

The study started in 2011-01, reached primary completion in 2016-03, and is listed as completed. Phase is recorded as not applicable. The registry record does not report the randomization unit, the number of sites, the allocation sequence, concealment procedures, or details of how external facilitation was delivered. It also does not give a full account of the usual-care condition.

The registry explicitly identifies the design as cluster-randomized in its analysis note. This matters because assignment and outcomes are related to sites, while the measured proportions concern patients within sites. Measurements from the same site may be correlated; treating every patient-level record as statistically independent would fail to reflect that structure. The analysis note therefore indicates that only the sponsor's analysis is shown rather than a simple individual-level re-analysis.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Participants were continuously entered into the study when starting a new antipsychotic treatment and were assessed up to 120 days later, thus the number of participants within each period is independent of one another and for some participants where the 120 days spanned two periods they are counted in more than one period.

Pre-implementation

Implementation

Sustainability

Analysis population for the primary outcome, as defined in the registry: All patients who met inclusion criteria for this measure. Note that total for Outcome 1 is less than overall total N because patients were included for each measure independently if they met criteria any time in the 6-month implementation phases. Slightly more patients met criteria for the "follow-up monitoring" measures than "baseline monitoring."

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: For each outcome measure, patients are selected when a new antipsychotic medication is prescribed. Some patients appear in more than one period (pre-implementation, implementation, sustainability); baseline characteristics are presented for the total sample (N=20,101) with rows for each period.

6. Primary Endpoint

The registry lists the primary outcome as Change in Site-level Rates of Weight Monitoring at Baseline (Within 30 Days of a New Antipsychotic Prescription), measured over Change in monitoring rates will be measured monthly through 6-month pre-implementation, implementation, and sustainability phases.

The primary endpoint concerns site-level rates of baseline weight monitoring, defined as weight recorded in the electronic health record for patients due for monitoring within 30 days of a new antipsychotic prescription. For each monthly observation, the measure is the proportion of eligible patients at a site with a recorded weight. The timeframe describes monthly measurement through pre-implementation, implementation, and sustainability phases.

Weight recording is a process measure: it assesses whether monitoring was documented, not whether patients experienced a particular change in weight or another metabolic outcome. It is used in this study because the stated purpose is to test implementation of recommendations for assessing and managing metabolic effects. A recorded measurement can indicate that a monitoring step occurred, but by itself it does not show what was done in response, whether the recorded value was accurate, or whether metabolic health changed.

The result is reported as an odds ratio of 1.19 with a 95% confidence interval from 1.02 to 1.38. An odds ratio compares the odds of an event between two groups, but the registry record gives coded group identifiers without mapping them to the named arms or identifying the numerator group. The estimate therefore cannot be translated from this record into a statement about which arm had higher monitoring odds. The record also does not report the underlying group-specific proportions or a p-value.

The population contributing to this endpoint is measure-specific. The registry notes that patients were included for each measure independently if they met criteria during the implementation phases, and that the total for this outcome is less than the overall total N. The truncated population description does not allow further detail about eligibility or the exact analysis population.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a sponsor analysis for a cluster-randomized, parallel-group study. The result is expressed as an odds ratio with a 95% confidence interval. The registry does not report the statistical model or enough methodological detail to reconstruct the analysis. The general explanations below describe the meaning of the statistical terms, not unreported features of the study's analysis.

What does an odds ratio describe?

General statistical explanation: Odds are the probability of an event divided by the probability that the event does not occur. An odds ratio compares odds between two groups. A value of 1 represents equal odds under the specified comparison; values on either side indicate different odds, with direction depending on which group is placed in the numerator. Because the registry does not map the coded groups to the named arms or identify their order, the reported value cannot be assigned confidently to intervention versus comparison.

An odds ratio is not the same as a difference in proportions or a ratio of probabilities. Its magnitude does not by itself reveal the absolute event rates. Interpreting the result in practical terms would require group-specific rates, along with a clear group mapping. Those details are not reported in the registry record.

How should the confidence interval be read?

General statistical explanation: A confidence interval conveys uncertainty around an estimate under the assumptions of the analysis. The reported 95% confidence interval runs from 1.02 to 1.38 for the odds ratio of 1.19. A confidence interval is not a guarantee that the true value lies inside its endpoints, nor does it account for bias from design or measurement problems. Its interpretation depends on the analysis model and sampling assumptions.

The registry does not report a p-value for this result. A confidence interval and a p-value are related ways of presenting statistical uncertainty in many analyses, but the absence of a reported p-value should not be filled in with a newly calculated value. The record also does not provide model specification, standard errors, or enough data for an independent calculation.

Why does clustering matter?

General statistical explanation: In a cluster-randomized study, groups such as sites, rather than individual patients, are assigned to study conditions. Patients observed at the same site may have more similar outcomes than patients from different sites because they share settings, practices, and implementation processes. This within-site similarity is called intracluster correlation. Ignoring it can make uncertainty appear smaller than it is because the analysis effectively treats related observations as more independent than they are.

The registry itself flags the cluster-randomized design and says a simple individual-level re-analysis would be misleading. It says only the sponsor's own analysis is shown. The registry record does not name the statistical method used to account for clustering, explain how repeated monthly measurements were handled, or describe adjustments for other factors. Those omissions limit independent appraisal of the reported interval.

What is the role of repeated monthly measurement?

General statistical explanation: Repeated observations can track how an outcome changes over time and can distinguish patterns across study phases. However, observations over time may be correlated within a site, and sometimes within the same person. A suitable analysis needs to reflect the dependence structure relevant to the design. The record describes monthly site-level observations across phases but does not report the exact model, how time was represented, or whether the analysis tested a change in trend, a difference at a particular time, or another contrast. The stated primary endpoint title refers to change in site-level rates, while the posted result gives one odds ratio; the relationship between that single estimate and the repeated measurements is not further explained.

What does a process measure capture?

General statistical explanation: A process measure records whether a care or monitoring step occurred. It can address implementation of recommendations, but it is not interchangeable with a clinical outcome. In this study, the primary endpoint is a recorded weight among patients due for monitoring. It does not directly quantify weight change, metabolic status, treatment decisions, or health consequences. Documentation may also differ from whether a measurement was actually taken, although the registry defines the measure through electronic health record recording and does not discuss validation.

How should planning illustrations be interpreted?

General statistical explanation: A sample-size calculation based on an observed effect is a planning illustration. It asks what sample size might be needed under stated assumptions to detect a chosen effect, not how much power the completed study actually had. The registry describes illustration assumptions including a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects defined as fractions of the observed effect. It does not provide calculated sample-size values in the record, and those assumptions are not reported as the original trial's design justification. The record does not report a sample-size justification, interim analyses, or a full statistical analysis plan.

10. Analyses Reported by the Sponsor

The sponsor's reported primary analysis is an odds ratio of 1.19 with a 95% confidence interval from 1.02 to 1.38. The registry's analysis note says the study was cluster-randomized and that a simple individual-level re-analysis would be misleading; it therefore presents only the sponsor's own analysis. No independent recalculation is reported in the registry record, and the analysis details needed to reproduce the estimate are not given.

For that reason, the sponsor estimate should be read as the available analysis rather than replaced by a calculation that treats patient observations as independent. The confidence interval communicates uncertainty under the sponsor analysis, but the record does not identify the model or explain how site clustering and monthly measurements were handled. The coded group identifiers are not mapped to the intervention and comparison arms, so the direction of the contrast cannot be assigned from the posted result alone.

The registry record does not report group-specific monitoring proportions or a p-value. Nor does it provide sufficient detail for an independent check of the estimate. The appropriate reading is therefore limited to the reported odds ratio and interval, together with the design warning and the missing group mapping. No additional statistical conclusion should be inferred from a recalculation that is not reported.

11. Multiplicity, Missing Data and Analysis Populations

The record lists one primary outcome concerning change in site-level baseline weight-monitoring rates and two secondary outcomes: change in site-level follow-up weight-monitoring rates, and change in management for obesity or weight gain after a recorded 5% gain in body weight. It also reports that the record contains three outcome measures and two analyses. The available information does not give results for the secondary measures or explain how the count of outcome measures relates to the listed outcomes.

Multiple outcomes and repeated measurements raise a general statistical issue: when many hypotheses are examined, the chance of at least one apparently notable result can increase unless the analysis plan accounts for multiplicity. The registry record does not report whether any multiplicity adjustment was planned or used, how outcomes were prioritized beyond identifying a primary measure, or whether the reported estimate was selected from multiple time points or models. It does not report interim analyses or a complete statistical analysis plan. These items should be described as not reported in the registry record, rather than assumed to have been absent or performed.

The analysis population is described only in part. The primary measure includes patients meeting the criteria for that measure, and the registry notes that each measure included patients independently if they met criteria during the implementation phases. It also states that the total for the primary outcome is less than the overall total N. The population description is truncated, so the exact eligibility rules, exclusions, and handling of patients with incomplete records are not reported in the registry record.

Missing-data methods are also not reported in the registry record. The definition uses electronic health record weight documentation, but the record does not explain how missing or unavailable weights were classified, whether missing data were distinguished from no monitoring, or whether any sensitivity analyses were conducted. Because the outcome is documentation-based, missing records and lack of a recorded weight may be difficult to distinguish from an unperformed measurement without further information.

The cluster-randomized design adds another population and analysis issue: the record does not explain the number of sites contributing to each analysis or how sites with different amounts of eligible information were handled. These omissions constrain judgments about representativeness and robustness, without establishing that any particular problem occurred.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The serious adverse-event figures list 0 serious events in the intervention group among 11335 at risk and 0 in the comparison group among 8766 at risk. The other adverse-event figures also list 0 events in each group, with the same at-risk counts. These figures describe the event counts and denominators recorded for those categories; they do not establish that no adverse events of any kind occurred.

The denominators are much larger than the listed enrollment of 12, and the registry record does not explain the unit or construction of these adverse-event denominators. They should not be interpreted as counts of randomized participants without that information. The record also does not report death counts: those entries are null, which is different from a reported count of zero.

The listed figures are not a complete safety profile. The registry does not provide event definitions, event-collection methods, follow-up details, or a description of how adverse events were attributed or compared. The empty list of top serious-event terms does not add detail about the event categories. A cautious reading is that the registry lists zero events in the reported categories and denominators, while important information about ascertainment and the death entries is not reported in the registry record.

13. What the Result Does, and Does Not, Mean

The primary result describes a statistical contrast in odds for a process measure: weight recorded in the electronic health record among patients due for baseline monitoring after a new antipsychotic prescription. The estimate is an odds ratio of 1.19, with a 95% confidence interval from 1.02 to 1.38. It is evidence about the posted analysis of that specified measure, not a direct estimate of weight change or a broader clinical outcome.

The result does not establish which named arm had higher odds because the registry presents coded group identifiers without mapping them to intervention and comparison or stating the numerator. It does not give the group-specific monitoring rates or an absolute difference. The odds ratio alone cannot show how common recorded monitoring was in either group, and the interval does not resolve the missing group labels.

The study's stated purpose was to test an implementation approach for recommendations concerning metabolic side effects of antipsychotic medications. The reported primary endpoint addresses weight monitoring; it does not itself establish whether diabetes or elevated lipids were assessed or managed, whether a recorded weight prompted a response, or whether patient health changed. The secondary outcomes listed in the record address follow-up monitoring and management after weight gain, but their results are not reported in the registry record.

The design also limits what can be inferred from the posted number. The study is cluster-randomized, and the registry warns that an individual-level re-analysis would be misleading. The record does not report the sponsor's model details, the mapping of analysis groups, or the complete analysis population. These gaps make it inappropriate to attach a more specific arm-level interpretation to the estimate.

Accordingly, the result should not be treated as proof that the implementation approach caused a particular improvement, nor as proof that it did not. It is a sponsor-reported odds ratio with a confidence interval for a site-level documentation measure, interpreted within a cluster-randomized study and with material details not reported in the registry record.

14. Why This Trial Matters Statistically

This trial illustrates why an implementation outcome and a clinical outcome must be kept distinct. The primary measure captures whether a weight was documented for patients due for monitoring. That is relevant to whether recommended monitoring is taking place, but it is not itself a measure of weight change, metabolic health, or the consequences of monitoring. Clear interpretation depends on keeping the endpoint aligned with the question it actually measures.

It also highlights the importance of the randomization unit. The registry identifies a cluster-randomized design and explicitly warns against a simple individual-level re-analysis. When assignment occurs by site, patient records within a site can share context and are not automatically independent observations. The analysis needs to respect that structure for its uncertainty estimates to be meaningful.

The result further shows why group labels and contrast definitions matter. An odds ratio is inherently directional: its interpretation depends on which group is in the numerator. Without a mapping from the coded groups to the named arms, a numerical estimate and confidence interval cannot support an unambiguous statement about which arm had higher odds. Reporting group-specific rates and a clear contrast would make the result easier to interpret in practical terms.

Finally, this record separates statistical reporting from design documentation. It gives an estimate and interval, but the model, missing-data handling, multiplicity approach, and detailed analysis population are not reported in the registry record. Those details are important for assessing precision, reproducibility, and the scope of conclusions. The lesson is not that the result is invalid, but that statistical interpretation depends on transparent descriptions of design and analysis as well as a numerical estimate.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record leaves several trial-specific details unresolved. Although it lists enrollment as 12, it does not identify the unit represented by that number or report the number of sites. That omission is particularly relevant in a cluster-randomized study, where the number and characteristics of clusters can shape both precision and generalizability. The record also does not describe the randomization unit or allocation procedures beyond labeling the study randomized and cluster-randomized.

The posted primary result lacks a mapping between coded groups and the named intervention and comparison arms. It therefore cannot be read as an arm-specific direction of effect. The registry also does not report group-specific outcome rates, the model specification, the handling of monthly observations, or the method used to account for clustering. The sponsor's estimate and interval cannot be independently reconstructed from the information reported.

The outcome is based on electronic health record documentation. The registry does not explain whether missing documentation was classified as no monitoring, how incomplete records were handled, or whether the recording measure was validated against measurements actually taken. The measure-specific eligibility description is incomplete, and the record notes that the outcome population differs from overall enrollment.

Results for the listed secondary outcomes are not reported in the registry record. The record also does not report multiplicity procedures, interim analyses, a sample-size justification, or a complete statistical analysis plan. These omissions prevent assessment of several choices that could affect interpretation; they are not evidence that the corresponding procedures were or were not undertaken.

Finally, the safety entries list zero events in the reported categories, but do not explain the large at-risk denominators, event ascertainment, or death counts. These reporting limitations call for a bounded interpretation of the posted results rather than a broader conclusion about safety or clinical outcomes.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Monitoring and Management for Metabolic Side Effects of Antipsychotics — NCT01875861.
• Registry record last updated 2019-02-12; status last verified 2019-02.
• Results first posted 2019-02-12.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

This completed randomized, parallel-group study examined an approach to implementing recommendations for assessing and managing metabolic effects of antipsychotic medications. The intervention was Evidence-Based Quality Improvement plus external facilitation; the comparison was usual care in the MIAMI Project context. The listed enrollment is 12, and the registry describes the study as cluster-randomized.

The primary measure concerns site-level baseline weight-monitoring rates, based on electronic health record documentation for patients due for monitoring within 30 days of a new antipsychotic prescription. The sponsor-reported odds ratio is 1.19, with a 95% confidence interval from 1.02 to 1.38. The coded groups are not mapped to the named arms, and the registry warns that individual-level re-analysis would be misleading.

Secondary outcomes concern follow-up monitoring and management after weight gain, but their results are not reported in the registry record. The record also leaves the analysis model, missing-data methods, multiplicity procedures, and several population details unspecified. The primary result should therefore be read as a limited statistical contrast for a site-level documentation measure, not as a direct measure of metabolic health or patient benefit.